Guides
9 min read
A/B Testing QR Code Landing Pages: Boost Conversions with Data
Learn how to A/B test QR code destinations to optimize conversion rates. Includes Thompson Sampling auto-optimization and best practices.
You printed 50,000 QR codes on product packaging. Which landing page converts better? With A/B testing, you split traffic between variants and let the data decide.
Why A/B Test QR Codes?
QR code scans represent high-intent traffic. Sending that traffic to an unoptimized page wastes potential conversions. A/B testing lets you:
- Compare two or more landing page designs
- Test different offers, CTAs, or pricing
- Measure real conversion rates per variant
- Automatically shift traffic to the winner
Setting Up an A/B Test
curl -X POST https://api.scanstack.dev/v2/qr/:id/ab-test \
-H "x-api-key: qr_your_key" \
-d '{
"auto_optimize": true,
"variants": [
{ "label": "Original", "target_url": "https://example.com/landing-a", "weight": 50 },
{ "label": "New Design", "target_url": "https://example.com/landing-b", "weight": 50 }
]
}'
Auto-Optimize with Thompson Sampling
Traditional A/B tests require you to manually pick a winner. ScanStack's auto-optimize uses Thompson Sampling, a Bayesian algorithm that:
- Starts with equal traffic distribution
- As conversion data arrives, automatically favors the better-performing variant
- Balances exploration with exploitation
- Converges on the best variant without manual intervention
Recording Conversions
// On your landing page's conversion event:
fetch('https://api.scanstack.dev/v2/ab-test/{testId}/convert/{variantId}', { method: 'POST' });
The testId and variantId are included in the scan webhook payload.
Best Practices
- Test one variable at a time — Change only the headline or only the CTA
- Wait for significance — Auto-optimize handles this for you
- Use clear conversion goals — Define success before you start
- Combine with conditional routing — A/B test different pages for mobile vs. desktop